I agree.
Everyone who is blaming the consumer for this is not just wrong, but absurdly so, not to mention (determinedly) confused about the difference between grease and anti-seize. They are completely different products, and grease is NOT NOT NOT anti-seize and will NOT perform the same function as well.
Saying grease is the tool for preventing rust seizing is like calling WD-40 a lubricant, which, unfortunately, is something a lot of people do. Anyone who has made the grease mistake may be helped by this information before it's too late.
Also, there are different types of anti-seize for this application. You want copper here, not nickel.
In case anyone is interested, I was wrong to call the problem with my mower "galling." I have learned that galling only happens with moving parts, and a shaft you move once a year or less is not a moving part. The proper term is "corrosion-induced adhesion," or you can just say "rust seizing."
It is also unfortunate to see that the turning front wheels are still being deliberately conflated with the non-turning rear wheels, even after painstaking and patient correction. The guy who bought the mower new screwed up the front wheels by ignoring the manual, but the manual does not mention maintenance for the rear wheel shafts, and the shafts, which are not moving parts, have no zerks or journals.
Hugo already knows these are different parts, and he has been told the manual does not mention greasing the rear wheels, which is the wrong remedy anyway. This has already been explained, so I don't think he should keep circling back as though it there were still some valid point to be made. I hope we don't have to beat this very dead horse again.
My bet: we will soon hear familiar hoofbeats.
As mentioned several times, JD makes a similar assembly with much looser shafts, and my JD wheels are about 35 years old with zero issues and no real maintenance. No greasing, anti-seize, or periodic disassembly.
Kubota had no excuse for doing it wrong, especially 25 years after JD, a feared competitor, made my JD deck. You don't sit back for 25 years and watch your competitors do smart things while you repeat the same obvious mistakes. Kubota pays people a lot of money to keep up with John Deere's doings. I don't have to email them and ask. Every competent big company watches the competition carefully.
Being too lazy to go outside and measure, I would guess the clearance on the JD is around 0.050" (radius, not diameter), so the shafts are very loose. And in perfect condition, on a mower that has seen lots of hard use. JD's design allows the adjustment pins to wear from the movement of the shafts, but the expensive shafts don't weld themselves to the deck, and they have not worn out.
Let me put it this way: if you remove the pins from a JD shaft after 35 years with no maintenance, it FALLS OUT.
No hydraulic press needed.
The JD design works perfectly because 1) the gap allows water to drain out so the parts can dry, 2) the parts can move a little, beating up any rust that tries to develop, and 3) the parts would have to have a gigantic layer of rust between them in order to seize together, and a layer like that can't form.
I am not a JD fan, but they were right about this.
For anyone who is interested in the solution to the anti-scalp wheel problem instead of the cause, which is obvious, a progress report. I solved the problem.
Drilling out the left-hand wheel shaft was not smart, although it was smarter than designing anti-scalp wheel shafts that cement themselves to decks. It took forever, and all it did was help me avoid a little cutting and welding, which is easier work.
Today, I cut the right-hand wheel support off the deck and put it in the 20-ton hydraulic press. I thought I would have to make a jig to hold the support, but--and this is good information for others who have this problem--a large impact socket will work just fine. I supported the part with a socket, and it did the job. You can only push the shaft so far, however, because the ratchet end of the socket is obstructed, so using a proper jig is better.
I thought using the socket would be easier than cutting metal stock and drilling it, but if I had had a nice piece of round steel bar, I would have used machine tools to make a jig.
Fortunately, my press is air-over-hydraulic, because I had to reverse the part maybe 20 times. The shaft was in there so tightly due to Kubota's booboo, the press did not knock it loose right away, and the shaft did not want to go very far before slowing down the press. On the first try, the shaft moved about 3/32", and over time, it got looser and moved farther.
I kept turning it and applying Deep Creep, and I cleaned it occasionally. Eventually, I got it to the point where it would move on the 3-ton arbor press, which is quicker to operate.
Having turned it maybe 15 times on the arbor press as well as wire-wheeling the exposed parts of the shaft with my industrial pedestal buffer, I got it to where I could move it by supporting the assembly in a vise and using a blacksmith's hammer and a punch. Turning it and whacking it over and over, I loosened it to the point where it finally fell out.
Let me illustrate just how bad Kubota's design is. I sanded the inside of the support, and I wire-wheeled the whole shaft and also made sure there were no burrs. It is STILL a tight slip fit, even with the plating worn down.
For those who don't know, a slip fit is a fit that is just loose enough not to be an interference fit or transition fit. An interference fit requires force because the internal part is actually larger than the ID of the external part. A transition fit is what you get when a shaft is the same size as a hole.
As proven above over and over, and by a bazillion John Deeres with shafts that didn't seize, a tight slip fit is totally unnecessary and inappropriate here. It means the part will start to stick the very instant it begins to rust, and rust is guaranteed in a carbon steel part in my particular climate, even indoors. Rust takes up more room than sound metal, so when an enclosed part rusts, the fit gets tighter. It's like the inner part is water freezing in a sealed pipe.
When you leave no room for expansion, like Kubota, any rust will lead to immense pressure, friction, and then seizing. As you can tell from my description, in this case, the seizing was too strong for tools like a sledge, a torch, and a three-foot pipe wrench. It was no joke. This shaft was fixed more stubbornly than typical press-fit shafts. A sledge will knock a pressed shaft out.
By the way, the reason heat works to release (some) rusted parts is that the heat expands both the inner and outer parts, crushing the rust between them and weakening the bond. This is why heat often works all by itself, even though you would expect to have to cool the inner part. A lot of people think you always have to cool the inner part, but that isn't true.
Not only did Kubota make the fit too tight; they put the wrong types of metal together. Kubota plated the shaft with a weak and ineffective layer of cadmium or something, but they left the plain old carbon steel of the tubing alone, just waiting for the inevitable condensation to creep down into it. Then they failed to mention maintenance in the manual.
They could have pressed a brass sleeve into the bore, not that this would be as good as simply making the shaft the correct size.
Cutting the support off was the way to go. There was no way to make this shaft come out without drilling it, using something like EDM or a cutting torch, or pressing it out. Pressing is easiest, and you can't do it with the part still on the mower unless you design a special tool. Three tons of pressure from the arbor press didn't faze it. I don't know if a 12-ton press would have worked. Twenty tons got it done.
I considered rigging up a large pulley puller to push it, but I suspected, correctly, that it would never provide enough force.
Now I just have to deburr and clean everything, weld the support back on, and prime and paint. If I can find an easy way to hold it in the mill, I may also open the tube up by maybe 0.020" to correct Kubota's obvious error.
For others who face this later, I used a 6" grinder with a Walter cutoff wheel, a cordless sawzall with a thin-metal blade I shortened, and an oscillating tool with these carbide blades I heard about from Project Farm: https://www.amazon.com/dp/B07YDMX328?ref_=ppx_hzsearch_conn_dt_b_fed_asin_title_2&th=1
I also played around with a Dremel and a hacksaw, but the above tools worked best. The carbide blades are a thousand times as good as the Ridgid blades I bought in the past.
Today or tomorrow, I'll prep the part and the mower for welding and painting, and then I'll put the other new parts on the mower, including the sun-damaged fender I broke with my genius method of removing it without taking the step off. By Monday, I should have a functional mower with 4 anti-scalp wheels.
